Principal Component Analysis(PCA) algorithm summary

Principal Component Analysis(PCA) algorithm summary

• mean normalization(ensure every feature has sero mean)
• Sigma = 1/m∑(xⁱ)(xⁱ)^T
• [U,S,V] = svd(Sigma) 
• u_{reduce =} u(:,1:K)
  
• Z = u_{reduce ' * X}

　　Pick smallest value of k for which 

　　∑^k_i=1 S_ii / ∑^i=m_i=1 S_{ii  >= 0.99}  (^{99% of variance retained})

　　

转载于:https://www.cnblogs.com/ChenAlong/p/5128135.html